1288 
other continually, because they do so at the end, when the values 
7, and M, are reached. 
In reality however during the time vr’ it is not the magnetization 
M indicated by the straight line OC that exists, but the magnetization 
M represented by the curve ABC. Consequently the action of the 
residual field on the magnetization M— M’ remains. This action gives 
rise to an impulse proportional to (J/—M’) +, if we denote by 
M—WM’ the mean value of the difference during the interval x’. If 
the negative values of J/-—M’ found between 0 and Q, surpass 
the positive ones between @Q and C, the mean difference M— M’ 
will liave the same direction as the remanent magnetization OA, 
which existed a moment before and which gave rise to effect 37. 
In this case the effects 3a and 35 have the same sign. In the case 
that has been supposed in the figure the positive values 1/—M’ 
are of more importance than the negative ones, so that here the 
two effects 8a and 35 have opposite signs *) ®). 
As it is desirable to make the third effect as small as possible, 
the experiment was arranged in such a manner, that the time T was 
short. This is why I used the pendulum. In order roughly to measure 
the time rt, the wires a and / were now connected, so that one 
and the same conductor was closed by the contact of the steel wire 
with « and by that with 6. The conductor in question was one of 
ay Here we may remark, that effect 3a@ would be preceded by an effect 3b of 
the same direction as 3a itself, if the disappearing of the current took an appre- 
ciable time. This follows from the circumstance that over its whole length FGA lies 
beneath the straight line OF. 
2) (u—M"’) or [MM dt over the time +’ might be calculated, if we 
knew, not only M as a function of 7 (i.e. the exact form of the curve in figure 2) 
but also for each value of 7, the number of lines of induction N enclosed by the 
windings, a number that might be represented by a curve similar to that in the 
figure Indeed, if 7 is the resistance of the circuit and # the electromotoric force 
dN 
=  — j 
dt 
or, since 
ri li 
(i : aN 
PIL — Oe 
ri her 
so that we may write for | (MM) dt 
1 (MM 
fi asstranied JN 
bet Tt 
The value of this expression might be derived from sufficiently accurate graphical 
representations. 
